[NET] Changed behavior in deprecated Base64.decode.

With 3.11.x the behavior for Base64.decode changed, previously it would seamlessly handle both urlSafe and standard modes.

As still documented in the constructor Base64(int lineLength, byte[] lineSeparator, final boolean urlSafe)

"@param urlSafe Instead of emitting '+' and '/' we emit '-' and '_' respectively. urlSafe is only applied to encode operations. Decoding seamlessly handles both modes."



The following now throws IllegalArgumentException: Illegal base64 character 2d

boolean urlSafe = false;
String encoded = "o6wBF2p1dwkzJbXY+vUaAg==";
final byte[] bytesToEncode = new Base64(0, new byte[] {}, false).decode(encoded);
urlSafe = true;
String encodedUrlSafe = new Base64(0, new byte[] {}, true).encodeToString(bytesToEncode);

final byte[] decoded = new Base64(0, new byte[] {}, false).decode(encodedUrlSafe);

Revisions before this commit still have the expected behavior: https://github.com/apache/commons-net/commit/deb24bf22b1f75e4dc1bc981f2bcd37f5c407ceb


Although Base64 is deprecated, its behavior should not change.

Thanks, 
Alexander